Impact of HPV Infection on the Immune System in Oropharyngeal and Non-Oropharyngeal Squamous Cell Carcinoma: A Systematic Review

Objectives: To review the current knowledge regarding the involvement of human papilloma virus (HPV) infection and the immune system in the development of head and neck squamous cell carcinoma (HNSCC). Methods: An electronic literature search was conducted to identify articles published between 1990 and 2019 pertaining to tumor-infiltrating immune cells (TICs) in HNSCC using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Issues of clinical relevance, including tumor location, the number of tumor samples, the inclusion of additional specimens (dysplastic or normal mucosa), tumor size, methods used for HPV detection, relationship between antigen expression and patient characteristics (age, gender, smoking, alcohol consumption, etc.), and prognostic data (overall survival (OS) and recurrence-free survival (RFS)) were assessed by four blinded investigators. Results: The search identified 335 relevant studies, of which 41 met the inclusion criteria. Of these, 7 studies focused on the peripheral blood immune cell concentration in patients with HNSCC according to HPV status, and 36 studies investigated TICs in the intraepithelial and/or stromal compartment(s) according to HPV status. The immune cells studied were CD8+ T cells (N = 19), CD4+ T cells (N = 7), regulatory T cells (Tregs, N = 15), macrophages (N = 13), myeloid-derived suppressor cells (MDSCs, N = 4), and Langerhans cells (LCs, N = 2). Conclusions: Irrespective of tumor location, CD8+ and CD4+ T cells appear to play a key role in the development of HPV−related HNSCC, and their infiltration is likely associated with a significant impact on OS and RFS. To date, the roles and prognostic value of Tregs, macrophages, DCs and MDSCs remain unclear.


Introduction
Head and neck squamous cell carcinoma (HNSCC) is the sixth most common cancer in men and the eighth most common cancer in women, accounting for over 600,000 new cases per year worldwide [1,2].

Materials and Methods
The criteria for considering studies for the systematic review were based on the population, intervention, comparison, and outcome (PICO) framework [16].
Types of studies: Prospective or retrospective clinical trials published in peer-reviewed journals were included in this review. Studies were included if they explored TICs in HNSCC, including oropharyngeal, laryngeal, hypopharyngeal and oral squamous cell carcinoma. We included studies published in English and French.
Participants and inclusion/exclusion criteria: Papers were included in the analysis if they clearly reported TICs in excised HNSCC samples or pretherapeutic biopsies from patients who were treated with conventional treatment (i.e., surgery, chemoradiation, chemotherapy, or immunotherapy). Papers that examined the recruitment of immune cells through blood analyses were also included to improve the understanding of the tumor immune microenvironment. HPV detection was required to have been performed through DNA or p16 analysis of the tumor samples. Studies focused on non-HNSCC were excluded.
Outcomes: The first outcome was the study of TICs in the intraepithelial (or intratumoral) and stromal compartments in HNSCC according to HPV status. To better understand the involvement of the immune system in the development of HPV−induced HNSCC, investigations into the expression of both cytokines and checkpoint proteins involved in tumor development were performed. Table 1 summarizes the immune cells included in this systematic review [17][18][19][20][21][22][23][24][25]. The second outcome was the study of tumor characteristics (tumor size, node, metastasis, and histopathology) and/or prognostic data (overall survival (OS) and recurrence-free survival (RFS)) in relation to TICs and HPV status.
Intervention and comparison: In the case of the study of the prognostic value of TIC, the authors were required to have treated their patients with conventional surgical or conservative treatments.

Search Strategy
Four authors (JRL, IS, TV, and QM) conducted searches in PubMed, Cochrane, and Scopus databases to identify articles published between January 1990 and June 2019 pertaining to the relationship between HPV infection and TICs in the development of HNSCC. Clinical studies were screened if they had database abstracts, available full texts or titles that referred to these conditions. The following keywords were used: 'HPV', 'cancer', 'carcinoma', 'head', 'neck', 'immune', and 'cell'. Final article selection was determined by the four authors, who provided a critical analysis of the publication content. The review was conducted according to the PRISMA checklist [26]. Institutional review board approval was not required. Number of CD8+ T cells increases throughout tumor progression. [9,13,18,23] 2. Recruitment of Treg lymphocytes. The infiltration CD8+ T cells is usually associated with a better 3. Upregulation of PD-L1 on tumor cells through an interferon (IFN)-γ-dependent host response to the tumor.
Manner. Higher CD8+ T cell infiltration is associated with better OS.

CD4± T cells CD4
The role of CD4+ helper T cells is unclear because a wide range of CD4+ cell subsets The high infiltration of CD4+ T cells is associated with controversial results. [13,18,23] with different functions exists. Some authors reported that in HPV−negative patient cohorts, a high level of CD4+ 1. Th-1 lymphocytes may activate cytotoxic lymphocytes.
TILs was associated with better OS and DFS, while contradictory 2. Th-2 lymphocytes stimulate humoral immunity and activate eosinophils. results were found for HPV+ patients.
3. Th-17 lymphocytes may have a Th-1 phenotype in the tumor microenvironment.
The role of Th-17 in HNSCC is still controversial. showed that an increased level of Tregs is linked with a worse prognosis. Others reported that high Treg counts are associated with a better prognosis. Immune escape is achieved by producing IL-10 and TGF-β and by consuming IL-2.
Dendritic cells Immature: CD1a 1. Presentation of tumor-associated antigens to the immune system (MHC class I & II) DC infiltration increases throughout tumor progression.
of Tregs, the silencing of differentiated antigen-specific T cell tolerance and the differentiation of naïve CD4+ T cells into Tregs.  (Lu).

Epidemiological Characteristics and Outcomes
An analysis of the locations of tumors, the number of tumor samples, the inclusion of additional specimens (i.e., dysplastic tissues or normal mucosa), tumor size, the relationship between antigen expression and patient characteristics (age, gender, smoking, and alcohol consumption), and prognostic data (OS and RFS) was performed. The method used for HPV detection was carefully analyzed to provide a methodologically critical analysis of the included studies.

Study Characteristics
The initial screening identified 335 studies, of which 41 met our inclusion criteria ( Figure 1). A total of 7 studies investigated the peripheral blood immune cell concentration in patients with HNSCC according to HPV status ( Table 2) [27][28][29][30][31][32][33]. Three studies focused precisely on the evolution of the peripheral blood immune cell concentration between pre-and posttreatment (surgery or chemoradiation) [28,29,31]. A total of 36 studies examined TICs in the intraepithelial and/or stromal compartment(s) according to HPV status ( Figure 1). Of the papers that investigated several different immune cell categories, some focused on oropharyngeal SCCs, whereas others focused on HNSCCs, including    A total of 36 studies examined TICs in the intraepithelial and/or stromal compartment(s) according to HPV status ( Figure 1). Of the papers that investigated several different immune cell categories, some focused on oropharyngeal SCCs, whereas others focused on HNSCCs, including oropharyngeal and other anatomical localizations. The studies are shown in Tables 3-6 [12,15,24,.

HPV and the Peripheral Blood Concentrations of Immune Cells
Because some immune cells are derived from bone marrow and are transported to tissues through the circulation, the peripheral blood number of immune cells (PBNI) was explored in 7 studies. Overall, it seems that patients with HNSCC have a higher PBNI than healthy individuals [28,29,32,33]. The role of HPV status is still unclear because only one study reported significant differences between the PBNI in patients with HPV+ and HPV− HNSCC [30]. Two authors found a high concentration of circulating CD8+ T cells against the E6 and E7 proteins in the majority of HPV+ patients with oropharyngeal SCC [29,31]. However, these observations were not supported by Heusinkveld et al., who included patients with oropharyngeal and non-oropharyngeal SCC [27]. According to Lukesova et al., natural killer cells are another type of cell that can be increased in patients with HPV+ oral or oropharyngeal SCC [30]. The other types of immune cells were less well studied, and it is difficult to pursue some lines of investigation.
According to two studies [28,29], the treatment type could have an impact on the PBNI, but the differences between these two studies (in terms of tumor location, types of treatment, and HPV detection methods) limited our comparison. The relationship between PBNI and OS was addressed in two studies [30,31]. Masterson et al. found that a high blood concentration of CD8+ T cells enhanced immunoreactivity to antigen E7, which was associated with improved OS in patients with oropharyngeal SCC [31]. In the same vein, Lukesova et al. demonstrated that patients with oral or oropharyngeal HPV+ SCC and a high Treg blood concentration had improved OS and RFS [30].

HPV and CD8+/CD4+ T Cell Tumor Infiltration
A total of 19 studies addressed CD8+ T cell infiltration (Table 3). Among the papers focusing on oropharyngeal SCC, the majority reported higher stromal or intraepithelial CD8+ T cell infiltration in HPV+ than in HPV− SCC [36][37][38]40,43,[45][46][47]51]. Only Wansom et al. did not find significant differences between these two types [34]. These results are relatively similar to those of studies that investigated HNSCC. Russel et al. and Badoual et al. observed increased CD8+ T cell infiltration in the intraepithelial compartment in HPV+ HNSCC [35,39], whereas when investigating laryngeal, oral, hypo-and oropharyngeal tumor samples, Balermpas et al. did not find significant differences between HPV+ and HPV− SCC [42]. In a second study, the same authors reported the occurrence of different patterns of CD8+ T cell infiltration according to the tumor location; CD8+ T cell infiltration was higher in HPV+ than in HPV− tumor samples only in oropharyngeal SCC [46]. Other authors reported an increase in infiltration in HPV+ HNSCC irrespective of the compartment [44,[48][49][50].                  The high CD8+ T cell infiltration in HPV+ tumors was associated with the high expression of checkpoint proteins, including PD-1 [44,48,52], LAG-3 [48], and Tim-3 [35], in HNSCC. The expression of PD-1 by HPV+ tumors was reported in the study by Kansy et al., who found a significant association between the CD8+ antigens CD8A and CD8B and the expression of PD-1 [52]. Partlova et al. also found that HPV+ tumors had higher expression of PD-1 mRNA than did HPV− HNSCC tumors [44]. The activity of CD8+ T cells is mediated by many cytokines, leading some authors to study the expression profile of TICs. Partlova et al. found that TICs in HPV+ SCC comprised IFN-g+ and IL-17+ CD8+ T cells [44], supporting the key role of these cytokines in the inflammatory reaction related to HPV. Overall, a high number of TICs (including CD8+ T cells) in HPV+ tumors was associated with the increased secretion of the following proinflammatory cytokines: CCL-17, CCL-21, IL-2, IL-4, IL-8, IL-10, IL-12, IL-17, IL-21, TNF-a, and IFN-g [39,44]. However, neither of these two studies demonstrated that cytokine expression was directly related to the level of CD8+ T cells.
A high level of CD8+ T cell infiltration was associated with improved OS and RFS in all studies focusing on HPV+ oropharyngeal SCC [34,36,37,40,41,51]. Moreover, Ward et al. found that patients with HPV+ oropharyngeal SCC with low levels of tumor-infiltrating lymphocytes had the same prognosis as those with HPV− oropharyngeal SCC [41]. Similar overall findings were observed in HNSCC for OS [42,46,48,50] and RFS [46,48,50]. One study reported that low CD8+ infiltration was associated with a high risk of metastases [46].

HPV and CD4+ T Cell Tumor Infiltration
Seven publications have reported findings regarding CD4+ T cell infiltration in HPV−induced SCC (Table 3). Overall, CD4+ T cell stromal infiltration was reported to be higher in HPV+ than HPV− oropharyngeal SCC [37,38,40,43]. In HNSCC, Balermpas et al. reported similar levels of CD4+ T cell infiltration in HPV+ and HPV− HNSCC [42]. Regarding the prognostic value of CD4+ T cell infiltration, Nordfors et al. did not find a significant association between the level of CD4+ infiltration and OS in patients with HPV+ and HPV− oropharyngeal SCC [40]. Similar results were found in the study by Balermpas et al., in which CD4+ expression was not associated with a good prognosis [42]. Regarding the subpopulation of CD4+ T cells, Krupar et al. studied the infiltration of Th17 cells in oropharyngeal SCCs. They found that there was a significantly lower percentage of Th17+ T cells in the intratumoral compartment of HPV+ patients.

Regulatory T Cell Infiltration
Fifteen studies have examined the tumor infiltration of Foxp3 Tregs according to HPV status, including 7 that were focused on oropharyngeal SCC (Figure 1, Table 4). Overall, a few studies reported that the degree of Foxp3 Treg infiltration did not differ according to HPV status in oropharyngeal SCC [34,45], but the majority of the authors observed increased infiltration in HPV+ oropharyngeal SCC, especially in the intratumoral compartment [36,38,43,53,54]. When analyzing HNSCC studies, Kindt et al. demonstrated that Foxp3 Treg infiltration increases with tumor progression; this increase is more important in HPV+ patients [24].
Four studies reported increased infiltration of Foxp3 T cells in the intratumoral [24,35,49] or stromal [39] compartment in HPV+ compared with HPV− HNSCC. Additionally, a study by Partlova et al. reported a slightly lower proportion of Treg cells in HPV+ HNSCC. These authors also found higher levels of cytokines and chemokines in HPV+ patients than in HPV− patients, but they did not report which cells produced these cytokines [44]. In 3 studies, the authors found that the infiltration of Foxp3 T cells was similar in HPV+ and HPV− SCC [15,33,42]. Two studies aimed to study the association between Treg infiltration and checkpoint protein expression [33,35]. Irrespective of HPV status, Badoual et al. and Lechner et al. found that the infiltration of Tregs was associated with higher levels of checkpoint protein expression (i.e., PD-1) in HNSCCs [33,35].
Regarding the prognostic value of Treg infiltration, the high infiltration of Foxp3 Tregs was associated with improved OS or RFS in both oropharyngeal cancer [34,36,53,54] and HNSCC [15,24,35,39]. Moreover, the results of the study by Kindt et al. suggested that the number of stromal Tregs is a strong prognostic factor that is independent of other risk factors, including tobacco and alcohol consumption and HPV status [24]. To date, only Balermpas et al. did not find a significant association between the degree of Treg infiltration and OS [42].

Macrophage Infiltration
From our literature search, we identified 5 and 8 publications that investigated macrophage infiltration in oropharyngeal and HNSCC according to HPV status, respectively (Figure 1 and Table 5). In 3 studies, CD68+ macrophages showed increased infiltration in HPV+ compared to HPV− oropharyngeal SCC [56][57][58][59]; only one publication did not corroborate this finding [34]. In both oropharyngeal and nonoropharyngeal SCC, CD68+ macrophage infiltration increases between dysplastic tissues and carcinoma; the macrophage infiltration density is higher in the advanced stages [14,57]. Yu et al. focused on the CD68+/CD163+ macrophage (M2) subpopulation in oral SCCs, and they did not find a significant difference between HPV+ and HPV− oral SCC according to tumor progression and HPV status [57]. However, the relationship between CD68+ macrophage infiltration and HPV status is more controversial in HNSCC. Indeed, 5 authors did not find a significant difference between HPV+ and HPV− HNSCC regarding CD68+ macrophage infiltration [39,44,55,57,62], while macrophage infiltration increased in the intratumoral compartment in HPV+ HNSCC in 3 studies [14,49,61]. Among these studies, Ou et al. reported a tendency toward a higher proportion of M2 macrophages in HPV− HNSCC [61].
The relationship between CD68+ macrophage infiltration and checkpoint protein expression has been examined in two studies [56,59]. On the one hand, Lyfor-Pike et al. reported that CD68+ macrophages expressed high levels of PD-1 [56]. On the other hand, Oguejiofor et al. found that HPV− oropharyngeal SCC had an increase in CD68+ PD-L1+ macrophages compared to HPV+ tumors; these tumors were associated with improved OS compared to HPV− oropharyngeal SCC with low CD68+ PD-L1+ infiltration [59]. Lee et al. found that the high infiltration of CD68+ macrophages was associated with poor OS and DSS [58], while Seminerio et al. demonstrated that the high intratumoral infiltration of CD68+ macrophages was linked with shorter OS in patients with HNSCC [14]. Finally, Welters et al. found that the increase in dendritic cell-like macrophage infiltration in HPV+ oropharyngeal SCC was correlated with improved OS and a low risk of lymph node metastases [60].
Regarding the expression of checkpoint proteins, Yu et al. found a positive correlation between the infiltration of MDSCs and PD-L1 expression [57]. For CD8+ T cells, Partlova et al. reported high proinflammatory cytokine expression in tumor samples, characterized by a high infiltration of MDSCs, but they did not demonstrate a potential association [44]. To date, no study has investigated the prognostic value of MDSC infiltration according to HPV status.

Langerhans Cells
Only 2 studies assessed LC infiltration according to HPV status (Table 6). In a cohort of 27 patients with oral SCC, Perreira et al. did not find a significant difference between HPV+ and HPV− oral SCC [63]. More recently, Kindt et al. observed that LC infiltration increased in HNSCC throughout tumor progression but decreased in the presence of HPV infection. Moreover, there was a significant association between the LC infiltration level and cT and tumor node status [24]. The LC infiltration level was positively associated with improved OS and RFS in HPV− but not HPV+ HNSCC.

Discussion and Perspectives
This systematic review emphasizes many lines of evidence and reveals uncertainties regarding the role of immune cells in the development of HPV−induced HNSCC. These findings are summarized in Table 7. Our analysis indicates that in both oropharyngeal and nonoropharyngeal SCC, HPV infection is associated with increased CD8+ T cell infiltration and PD-1 expression by CD8+ T cells and improved OS. CD8+ T cells interact with their environment through multiple cytokines, especially IFN-g and IL-2, -4, -8, -12 and -17 [39,44]. The increased expression of IL-17 in HPV+ HNSCC cells is associated with the infiltration of Th17 lymphocytes, resulting from the differentiation of CD4+ T cells. Thus, Krupar et al. reported an increased number of Th17 lymphocytes in TICs in HPV+ oropharyngeal SCC [43]. This observation makes particular sense according to studies that demonstrated increased CD4+ T cell infiltration in HPV+ SCC [37,38,40,43]. Because recent studies have suggested that lymphocyte plasticity can occur during HNSCC development, which is characterized by Th1 phenotype expression in Th17 cells [65,66], these findings could support a potential relationship between CD8+ and CD4+ T cells in HPV−induced SCC. CD4+ T cells could be converted into Th17 cells, potentiating the cytotoxic effects of CD8+ T cells against HPV−induced SCC antigens. According to several studies, the development of TICs, including CD8+ T cells, is associated with an increase in the detectable PBNI [28,29,32,33]. In this respect, and considering the data of Parikh et al., it is probable that the detection of CD8+ T cells against E6 or E7 HPV proteins could be used in future studies to better characterize tumor immunogenicity and, as recently suggested, the response to some treatments (i.e., immunotherapy and chemoradiation) [29,31]. However, scholars should remain prudent because the usual method of assessing the PBNI (blood sampling) is performed at a single time point and does not usually consider the variation in the blood concentrations of immune cells due to other factors related to the circadian rhythm or other external causes.
Macrophages have also been extensively studied in relation to HPV status. However, only a few studies have performed coimmunostaining to identify the M1 and M2 phenotypes [61]. M2 macrophages are involved in the enhancement of immunosuppression through the stimulation of Tregs and the secretion of TGF-b, TNF-a, and IL-10, leading to the creation of a favorable tumor microenvironment. Recently, Ou et al. indicated that CD68+ macrophage infiltration in HPV−SCC may consist mostly of M2 macrophages [61]. In HPV−HNSCC, the increased proportion of M2 cells could be one factor underlying the improvement in OS.   Because Foxp3 Tregs inhibit the activity of CD8+ T cells, these cells have been examined in many studies, and their relationship with the tumor microenvironment in HPV−induced SCC remains ambiguous. Indeed, in contrast with expectations, most studies reported that an increase in Treg infiltration, especially in HPV+ SCC, was associated with improved OS [15,24,[34][35][36]39,53,54]. Irrespective of HPV status, some authors have reported similar findings in HNSCC [66,67]. In fact, the high infiltration of Foxp3 Tregs may inhibit the protumoral effects of inflammatory immune cells and may function as favorable prognostic markers at some tumor sites, whereas at other tumor sites, Treg infiltration may be linked to poor OS due to their conventional regulatory function [41,68,69]. The activation of Tregs may be associated with MDSC infiltration because DC subpopulations are known to stimulate Tregs in HNSCC [69]. However, we cannot advance any hypotheses about the potential role of DC in the development of HPV−induced SCC because of the low number of studies on this topic in the current literature. Future studies that aim to investigate the role of Foxp3 Tregs in HPV−induced HNSCC should examine other immune cell populations, such as MDSCs.
This systematic review permitted the identification of many factors that should be taken into consideration in the analysis of results, the comparison of studies and the future establishment of immune models.
First, many authors have combined patients with oropharyngeal and non-oropharyngeal SCC into a single group. However, HNSCC includes malignancies that arise from functionally and anatomically distinct tumor sites with different characteristics. The most blatant example concerns oropharyngeal histology, as the Waldeyer ring is composed of preexisting lymphoid tissue that is characterized by a higher sensitivity to the immune response [49]. In that respect, oropharyngeal SCC is usually heavily infiltrated by lymphocytes, in contrast to other types of HNSCC such as oral carcinoma [49,69]. Additionally, recent data have shown the heterogeneous molecular and immunological tumor profile of HNSCC at different anatomical locations [70]. Differences in the prognostic value of TIC might reflect these different biological factors, making it likely that TICs exhibit different properties depending on the tumor site and the histological and molecular subtype.
In addition to the impact of the anatomical location of tumors, the consumption of tobacco and alcohol may also have a critical impact on the significance of TICs. Indeed, because cohorts of HPV+ patients are usually small, most authors do not consider these factors to be important in the development of TICs. However, it has been suggested that tobacco and alcohol consumption are capable of stimulating the mucosal recruitment of LCs, impacting the local immune response during SCC development [18]. In the same vein, Geng et al. demonstrated that tobacco smoking and, in particular, nicotine, are known to impair the responsiveness of T cells to antigenic stimulation, while other authors found that smoking is associated with a lower number of CD8+ T cells in tissue [46,71,72].
Second, the analysis of TICs could ideally consider both the stromal and intraepithelial compartments because the level of infiltration could vary substantially between these two compartments, leading to differences in prognostic value. Some authors did not specify the compartment used for analysis, which makes an analysis of the results difficult [40,41,44].
A third important point regarding the results analysis is the method used to detect HPV infection. In clinical practice, HPV infection is detected through p16 immunostaining in many centers. Some authors have used this approach to compare data from HPV+ and HPV− SCC, while others have used PCR and other direct methods of DNA identification. However, it is well known that some tumors can be HPV+/p16− or HPV−/p16+, which can be related to the lack of specificity of p16 in identifying HPV infection. As demonstrated in two studies, the use of p16 immunostaining versus DNA detection may lead to differences in the results, biasing the conclusions of related studies [40,54].
This systematic review also showed that some immune cells are rarely studied, such as eosinophils or natural killer cells (NK cells). However, NK cells have been identified as immune cells that may directly kill both HPV+ and HPV− HNSCC tumor cells [72,73]. This anti-tumor activity of NK cells could be significantly enhanced by cetuximab or avelumab, particularly in cells with higher baseline EGFR or PD-L1 expression [73]. Clinically, irrespective of HPV status, the low number of tumor-infiltrating CD56+ NK cells is correlated with significantly decreased OS, distant metastasis-free survival and local progression-free survival [17].
Finally, the studies that investigated the impact of HPV infection on OS included patients subject to different treatment methods, including surgery, chemoradiation, radiotherapy and immunotherapy. These different therapeutic methods may significantly impact the OS of patients irrespective of HPV status. This point must be considered in analyses of OS in future large cohort studies.
In conclusion, current knowledge regarding the immune environment of HPV−induced HNSCC is not sufficient for the establishment of a clear pathophysiological model. Although some evidence exists in terms of the roles of CD4+ and CD8+ T cells and their related impact on OS or RFS, many uncertainties persist regarding the role and prognostic value of Tregs, macrophages, DCs and other uninvestigated cells. The poor quality and the low number of available studies, the small number of HPV+ patients included in these studies and the lack of consideration of cofactors that can impact the TIC composition may explain the current inability to establish an immunological model that can better predict the prognosis of HPV−induced oropharyngeal SCC. Future studies are needed to understand the complex interaction between tumors and their immune environment. These studies should carefully consider a rigorous methodological approach for HPV detection and should include a large number of patients with well-defined tumor locations. A large panel of immune cells and the use of specific coimmunostaining should be considered in future work in order to establish a precise immunological overview of HPV+ and HPV− SCC.

Conflicts of Interest:
The authors declare no conflict of interest.